DATASOURCE Procedure

The Bureau of Economic Analysis, U.S. Department of Commerce, supplies national income, product accounting, and various other macroeconomic data at the regional, national, and international levels in the form of data files with various formats and on various media.

The following BEA data file types are supported.

The Bureau of Labor Statistics, U.S. Department of Labor, compiles and distributes data on employment, expenditures, prices, productivity, injuries and illnesses, and wages.

The following BLS file types are supported.

The DRIBASIC (formerly CITIBASE) database contains economic and financial indicators of the U.S. and international economies gathered from various government and private sources by DRI/McGraw-Hill, Inc. There are over 8000 yearly, quarterly, monthly, weekly, and daily time series.

Global Insight, formerly DRI/McGraw-Hill, distributes Basic Economic data files on various media. Old DRIDDS data files can be read by DATASOURCE using the DRIDDS filetype.

The following DRI file types are supported.

COMPUSTAT data files, distributed by Standard & Poor’s Compustat Services, Inc., consist of a collection of financial, statistical, and market information covering several thousand industrial and nonindustrial companies. Data are available in both an IBM 360/370 format and a "Universal Character" format, both of which further subdivide into annual and quarterly formats.

The BY variables are used to select individual companies or a group of companies. Individual companies can be selected by their unique six-digit CUSIP issuer code (CNUM). A number of specific groups of companies can be extracted by the following key fields:

Series names are internally constructed from the data array names documented in the COMPUSTAT manual. Each column of data array is treated as a SAS variable. The names of these variables are generated by concatenating the corresponding column numbers to the array name.

Missing values use four codes. Missing code ’.C’ represents a combined figure where the data item has been combined into another data item, ’.I’ reports an insignificant figure, ’.S’ represents a semi-annual figure in the second and fourth quarters, ’.A’ represents an annual figure in the fourth quarter, and ’.’ indicates that the data item is not available. The missing codes ’.C’ and ’.I’ are not used for Aggregate or Prices, Dividends, and Earnings (PDE) files. The missing codes ’.S’ and ’.A’ are used only on the Industrial Quarterly File and not on the Aggregate Quarterly, Business Information, or PDE files.

The Center for Research in Security Prices provides comprehensive security price data through two primary stock files, the NYSE/AMEX file and the NASDAQ file. These files contain master and return components, available separately or combined. CRSP stock files are further differentiated by the frequency at which prices and returns are reported, daily or monthly. Both daily and monthly files contain annual data fields.

CRSP data files are distributed in CRSPAccess format. For more information about accessing your CRSPAccess database, see Chapter 46, SASECRSP Interface Engine. You can convert your CRSPAccess data to binary format (SFA format) by using the CRSP-supplied utility (STK_DUMP_BIN). Use the DATASOURCE procedure for SFA format access and use SASECRSP Interface for CRSPAccess.

CRSP stock data (in SFA format) are provided in two files, a main data file containing security information and a calendar/indices file containing a list of trading dates and market information associated with those trading dates.

The file types for CRSP stock files are constructed by concatenating CRSP with a D or M to indicate the frequency of data, followed by B, C, or I to indicate file formats. B is for host binary, C is for character, and I is for IBM binary formats. The last character in the file type indicates if you are reading the Calendar/Indices file (I), or if you are extracting the security (S) or annual data (A). For example, the file type for the daily NYSE/AMEX combined data in IBM binary format is CRSPDIS. Its calendar/indices file can be read by CRSPDII, and its annual data can be extracted by CRSPDIA.

Starting in 1995, binary data used split records (RICFAC=2), so the 1995 filetypes (CR95*) should be used for 1995 and 1996 binary data. If you use utility routines supplied by CRSP to convert a character format file to a binary format file on a given host, then you need to use host binary file types (RIDFAC=1) to read those files in. Note that you cannot do the conversion on one host and transfer and read the file on another host.

If you are using the CRSPAccess Database, you will need to use the utility routine (stk_dump_bin) supplied by CRSP to generate the UNIX binary format of the data. You can access the UNIX (or SUN) binary data by using PROC DATASOURCE with the CRSPDUS for daily or CRSPMUS for monthly stock data.

For the four-digit year data, use the Y2K-compliant filetypes for that data type.

For CRSP file types, the INFILE= option must be of the form

INFILE=( calfile security1 < security2 \ldots > )

where calfile is the fileref assigned to the calendar/indices file, and security1 < security2 …> are the filerefs given to the security files, in the order in which they should be read.

The DATASOURCE procedure provides access to FAME Information Services databases for UNIX-based systems only. For information about a more flexible FAME database access, see the section "The SASEFAME Interface Engine" in Chapter 47, SASEFAME Interface Engine.

The DATASOURCE interface to FAME requires a component supplied by FAME Information Services, Inc. Once this FAME component is installed on your system, you can use the DATASOURCE procedure to extract data from your FAME databases by giving the following specifications.

Specify FILETYPE=FAME in the PROC DATASOURCE statement and give the FAME database name to access with a DBNAME=’fame-database ’ option. The character string you specify in the DBNAME= option is passed through to FAME; specify the value of this option as you would in accessing the database from within FAME software.

Specify the output SAS data set to be created, the frequency of the series to be extracted, and other usual DATASOURCE procedure options as appropriate.

Specify the time range to extract with a RANGE statement. The RANGE statement is required when extracting series from FAME databases.

Name the FAME series to be extracted with a KEEP statement. The items in the KEEP statement are passed through to FAME software; therefore, you can use any valid FAME expression to specify the series to be extracted. Enclose in quotes any FAME series name or expression that is not a valid SAS name.

Name the SAS variable names you want to use for the extracted series in a RENAME statement. Give the FAME series name or expression (in quotes if needed) followed by an equal sign and the SAS name. The RENAME statement is not required; however, if the FAME series name is not a valid SAS variable name, the DATASOURCE procedure will construct a SAS name by translating and truncating the FAME series name. This process might not produce the desired name for the variable in the output SAS data set, so a rename statement could be used to produce a more appropriate variable name. The VALIDVARNAME=ANY option in your SAS options statement can be used to allow special characters in the SAS variable name.

For an alternative solution to PROC DATASOURCE’s access to FAME, see the section "The SASEFAME Interface Engine" in Chapter 47, SASEFAME Interface Engine.

Haver Analytics offers a broad range of economic, financial, and industrial data for the United States and other countries. For information about accessing your HAVER DLX database, see the section "The SASEHAVR Interface Engine" in Chapter 49, SASEHAVR Interface Engine. SASEHAVR is supported on most Windows environments. Use the DATASOURCE procedure for serial access of your data. The format of Haver Analytics data files is similar to the CITIBASE/DRIBASIC formats.

The International Monetary Fund’s Economic Information System (EIS) offers subscriptions for their International Financial Statistics (IFS), Direction of Trade Statistics (DOTS), Balance of Payment Statistics (BOPS), and Government Finance Statistics (GFS) databases. The first three contain annual, quarterly, and monthly data, while the GFS file has only annual data.

PROC DATASOURCE supports only the packed format IMF data.

The Organization for Economic Cooperation and Development compiles and distributes statistical data, including National Accounts and Main Economic Indicators.